JPS593610A - Ipl retry processing system - Google Patents

Abstract

PURPOSE:To redress an intermediate fault and a fixed fault of either of the 1st and the 2nd buses by providing the 1st and the 2nd buses each with an IPL equipment-item-number setting means and a retry frequency setting means for specifying the retry frequency of IPL processing. CONSTITUTION:Equipment item numbers and the frequency of IPL processing to be retried are set in IPL equipment-item-number setting registers 7 (specifying the 1st access bus) and 8 (specifying the 2nd access bus), and a retry frequency register 9 from a display console 6 or floppy disk 13 in advance. If abnormality occurs during processing, a +1 circuit 12 is operated until a retry counter 14 goes up to the contents of the retry frequency register 9 and then error information is collected when the counted value becomes equal to the contents of the register 9. A substitute bus execution counter 11 counts up by one and a decision on whether execution from a substitute bus is completed or not is made.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technical Field of the Invention The present invention is an IPL retry processing system, in particular, an IPL retry processing system that combines initial program load (IPL) processing from a direct access device with duplicated access paths into a single IPL retry processing system. access·
Even if a path has a fixed failure, IPL can be automatically executed any number of times from other access paths, increasing system availability (2) Background and problems In general, access to the direct access device (DA8D) subsystem・Path duplication is provided for the purpose of improving performance and as a remedy against path failure.

FIG. 1 shows an example of path duplication for a DASD subsystem. In the figure, l is the central processing unit (CPUI, 2-1.2-2
is a channel device, 3.3-1.3-2 is a disk control device, and 4-1 to 4-3 are magnetic disk devices.

Duplication of access paths can be achieved, for example, when two channel devices 2-1, 2-2 are connected to the disk control device 3 as shown in FIG. (B) As shown, channel device 2-1 and disk control device 3-1
In some cases, the channel device 2-2 and the disk control device 3-2 are connected to each other. Access to the magnetic disk device 4-1 etc. is possible from both the first access path (PATH1) and the second access path (PATH2) in any case. The access path is usually determined by the machine number associated with the channel number, controller number, and unit number.

In order to start operating the system, it is first necessary to initialize various states for system operation. CPU
1 fetches and executes instructions in main memory sequentially, but
In particular, it is necessary to first load the instruction to be executed into main memory. This loading is the initial
This is called a program load (IPL) and is performed from a specific system resident volume mounted on a direct access device such as a magnetic disk device. After the initial settings of the system are completed, input/output to the magnetic disk device 4-1 etc. is performed using either of the above two access paths, since the input/output control section of the operating system is already prepared in the main memory. If one of them becomes a failure, input/output processing from another access path is automatically executed. However, IPL
In some cases, the above input/output control unit is not provided in the main memory, so conventionally, IPL is performed only on the specified main access path, and if it fails, the IPL is terminated without using the alternate bus. It was like that.

Therefore, duplication of access paths has no meaning at the time of IP'L, and especially in unmanned driving systems, if there is a failure in the main path, the system cannot be started up, so the loss is It was big.

(8) Object and Structure of the Invention The present invention aims to solve the above problems by making it possible to specify four access paths for IPL and the number of retries for IPL processing in advance, and automatically retrying the specified first access/
The purpose of this approach is to repeat retries from the path and from the second access path to enable IPL from the alternate path even when the main path has a fixed failure. Therefore, the IPL re-processing method of the present invention has a system resident volume on a direct access device with duplicated access paths, and has a startup processing power of a data processing system that performs IPL from the system resident volume. , an IPL specifying an IPL machine number for each of the first path and the second path of the access path.
machine number setting means; means for specifying the number of retries of the IPL process to be repeated when the IPL process is unsuccessful; and a first pass IPL machine instructed by the IPL machine number setting means; The present invention is characterized by comprising a try processing means for retrying the IPL processing as many times as instructed by the retry count setting means until the IPL processing is normally completed for each of the IPL machine numbers of the second pass and the IPL machine number of the second pass. Embodiments will be described below with reference to the drawings.

(4) Embodiment of the Invention FIG. 2 shows the configuration of an embodiment of the present invention, and FIG. 3 is a process explanatory diagram of the IJ) lie processing section according to the present invention.

In the figure, numerals 1 to 4-2 correspond to those in FIG. 1, 5 is a service processor (SVP), 6 is a display console, 7 and 8 are IPL machine number setting registers, 9 is a retry count register, and 10 is a IJ) Lie processing section, 11
12 is a plus-one circuit, 13 is a floppy disk, and 14 is a retry counter.

The service processor 5 is a central processing unit (CPU)
This is a subsystem that has 1llffi independent processors, and is capable of collectively monitoring, operating, and controlling the system constituted by the CPU 1 and the like.

It has devices such as a note display console 6 dedicated to the service processor 5 and a floppy disk 13. Other general functions and configurations of the service processor 5 are well known and will not be described here.

For example, assume that a system resident volume is mounted on the magnetic disk device 4-1, and that the IPL record of the system is provided in this system resident volume. This IPL record includes the operating
A program for loading the core of the system onto the main memory is stored, and when the IPL is operated, the IPL record is first read into the main memory.

Access to the magnetic disk device 4-1 can be performed via a first access path via the channel device 2-1 or a second access path via the channel device 2-2.

IPL machine'! The setting register 7 is set with an IPL machine number that specifies the first access path, and the IPL machine number setting register 8 is set with an IPL machine number that specifies the second access path.

The machine numbers are set in these IPI and machine number setting registers 7 and 8 by inputting commands from the display console 6. In addition, this content is
It is stored on disk 13, and when automatically executing IPL in an unmanned operation system, etc., it is stored on floppy disk 1.
3 and set.

The retry count register 9 is a register in which the number of retries to be made is set when input processing of an IPL record from the magnetic disk device 4-1 does not go well.

As in the case of I25 machine number setting register 7.8, it is set by input information from display console 6 or floppy disk 13. For example, access)
(If the failure on the access path is a fixed and permanent failure, input processing will not be recovered even by retrying from the same access path, but if the failure is an intermittent failure due to noise etc. It may be recovered by input processing from .

When the alternate path execution counter 11 is "0", input processing from the first access path is executed, when rlJ, input processing from the second access path is executed, and when it is "2", it is finished. This is a counter that indicates. Retry Kara/Ri1
4 is a counter that counts the number of retries, and the plus-one circuit 12 operates every time there is a retry. The retry processing unit IO actually executes the IPL retry, and processes as shown in FIG.

When an IPL request is issued to the service processor 5, the retry processing unit IO is activated.

The retry processing unit 10 first clears the alternate pass execution counter 11 to zero by the process 20 shown in FIG. Then, in step 21, the retry counter 14 is cleared to zero, and in step 22, it is checked whether the alternate pass execution counter 11 is zero. Since the initial value is zero, the first access path is selected in process 23, and the IPL machine number "XX" from the I25 machine number setting register 7 is
×” is set, and by processing 25, the channel device 2
-1, 0 that executes input/output commands to the magnetic disk device 4-1 via the disk control device 3;
Read the PL record to the main memory. Process 26 determines whether process 250 input/output processing was successful. If successful, control is transferred to the CPU 1 as a normal end.

In the case of abnormal termination, processing was previously stopped at that point, but
In the present invention, the processing is further continued as follows.

Through process 27, the plus-one circuit 12 is operated and the value of the retry counter 14 is incremented by one. Then, in step 28, it is determined whether the contents of the retry counter 14 have become equal to the contents of the retry count register 9. If they are not equal, which process 2
Return to step 2 and repeat the process. Therefore, if the process 25 does not end normally, the IPL process from the first access path is executed in the retry count register 9. will be retried for the value of .

If the judgments in process 28 are equal, it is regarded as a fixed failure,
Process 29 collects error information and stores it on the floppy disk.
The data is stored on the disk 13. Next, in process 30, the alternate pass execution counter 11 is incremented by +1. Then, in process 31, it is determined whether the execution from the alternate path has been completed. If the value of the alternate bus execution counter 11 is "2", it is determined that the process has ended. The substitution pass execution counter 11 initially changes from "O" to "1".
'', so it is not ``2''. Therefore, the process returns to step 21 and the process is repeated in the same manner. However, since the determination in process 22 is "NO," this time, in process 24, 1PLi number [YYYJ of the second access bus is read from the IPL and machine number setting register 8 and set, and the channel device 2-
Input/output to and from the magnetic disk device 4-1 is executed from the bus on the second side. Regarding this second access bus, if the second access bus does not terminate normally, it is retried the number of times set in the IJ) number of runs register 9 and error information is collected. is counted and amplified from "l" to "2".

Therefore, the process 31 determines that the process has ended with an error.

(5) As described in detail, according to the present invention, retrying an arbitrary number of times is possible.
It is automatically executed for each of the duplexed buses and can rescue intermittent failures and fixed failures of one access bus. Particularly in unmanned driving systems, etc., even if the above-mentioned failure occurs, the system can be started up without operator intervention, which is very effective.

[Brief explanation of the drawings] Figure 1 is an example of bus duplication of the DASD subsystem;
The figure shows the configuration of an embodiment of the present invention, and FIG.
-1 and 2-2 are channel devices, 3 is a disk controller, 4-1 is a magnetic disk device, 5 is a service processor, 6 is a display console, 7 and 8 are I
PL machine number setting register, 9 is retry count register, 1
0 represents a retry processing section. Patent applicant Fujitsu Ltd. Representative Patent Attorney Mori 1) Hiroshi (1 other person) Sai 11 yen] 1,300 yen

Claims (1)

[Claims] In a startup processing method for a data processing system that includes a system resident volume on a direct access device with duplicate access paths and performs IPL from the system resident volume, the first path and the second path of the access path are an IPL machine number setting means for specifying an IPL machine number for each of the IPL machine numbers, a number of retries setting means for specifying the number of retries of the IPL process to be repeated when the IPL process is unsuccessful, and an IPL machine number setting means for specifying the number of retries for the IPL process to be repeated when the IPL process is unsuccessful; For each of the IPL machine number of the first pass and the IPL machine number of the second pass, the IPL process is retried the number of times instructed by the retry count setting means until the IPL process is normally completed. ■PL retry processing method featuring the following features.